TWI478527B - Energy efficiency ethernet with asymmetric low power idle - Google Patents

Description

Energy-saving Ethernet network with asymmetric low power idle and method thereof

The present invention relates generally to an energy efficient Ethernet, and more particularly to an asymmetric low power idle energy efficient Ethernet. .

Energy costs continue to rise, and this trend has accelerated in recent years. In this case, various industries have become increasingly sensitive to these rising costs. One area that is attracting more and more attention is the IT infrastructure. Many companies are currently reviewing the power usage of their IT systems to determine if they can reduce energy costs. To this end, industries that focus on energy-efficient networks have emerged to address the rising cost of IT equipment in general (ie, PCs, displays, printers, servers, network devices, etc.).

One of the considerations in designing energy-efficient solutions is the use of network links. For example, many network links are typically idle between occasional bursts of data. The transmission of idle signals on the link wastes energy and increases the level of radiation. Therefore, the identification of these frequent low link utilization periods can provide opportunities for energy savings.

However, in other network links, the traffic profile may include regular or intermittent low frequency wide traffic with high frequency wide traffic bursts. Here, the identification of the low link utilization period is more difficult and the energy saving potential is lowered.

Traditionally, an energy efficiency control policy in a network device is an operational analysis of link utilization to determine whether to enter a low power idle mode to conserve power. Due to the two different ends of the link Data does not necessarily appear at the same time, so identifying the timing of entering a low-power idle mode can be difficult. Therefore, what is needed is a mechanism that maximizes energy savings when considering the asymmetry of link utilization.

The present invention is described in detail with reference to the particular embodiments illustrated in the drawings. The present invention is to be considered as illustrative and not restrictive

The present invention provides a method comprising: operating a network device in an active mode in which a physical layer device and a link partner are in a two communication direction via a network link in a network device Communicate at a transmission rate of 1 Gb/s; monitor the link utilization level of the first communication direction via the network link through the network device; and respond to monitoring to change the physical layer device from the startup mode to the asymmetric low power The idle mode, wherein the asymmetric low power idle mode supports the configuration of the physical layer device, wherein the first communication direction is in an unactivated state between the transmission of the periodic refresh signal, and the periodic refresh signal is configured to maintain the first communication The direction is synchronized, and the second communication direction opposite to the first communication direction continues to communicate at a transmission rate of 1 Gb/s.

The present invention also provides a method comprising: operating a network device in a boot mode in which the physical layer device supports and the link partner has a defined data transfer rate via the network link in the network device Two-way communication; and the response link uses level analysis to boot the physical layer device from The mode transitions to an asymmetric low power idle mode, wherein the asymmetric low power idle mode supports configuration of the physical layer device, wherein the first communication direction runs at a defined data transmission rate, and the second communication direction is periodic. The refresh signal is not activated between transmissions, and the periodic refresh signal is configured to maintain synchronization of the first communication direction, wherein the periodic refresh signal supports a logical channel having a maximum data rate lower than the defined data transmission rate.

The present invention provides a network device comprising: a transmitter configured to transmit to a link partner at a transmission rate of 1 Gb/s via a network link; a receiver configured to be 1 Gb/s via a network link The transmission rate is received from the link partner device; and the power saving control strategy is configured to analyze the link utilization level of the first communication direction via the network link connecting the network device and the link partner device, and the energy saving control strategy further Configured to respond to the analysis, transforming the network device into an asymmetric low-power idle mode, wherein the asymmetric low-power idle mode supports configuring the transmitter to be in an inactive state and to receive the receiver between the transmissions of the periodic refresh signal Configured to receive data at a 1 Gb/s transmission rate, the periodic refresh signal is configured to remain synchronized with the link partner device.

Various embodiments of the invention are discussed in detail below. Although specific implementations are discussed, it should be understood that this is done for illustrative purposes only. Those skilled in the relevant art will recognize that other components and configurations can be used without departing from the spirit and scope of the present invention.

When the traffic utilization of the network is not at its maximum capacity, the energy-saving Ethernet network tries to save energy. This is used to minimize performance impact while maximizing energy savings. Energy saving can be asymmetrically applied to a link providing an asymmetric low power idle mode, which supports configuration of a physical layer device, wherein the first direction of communication is between the periodic refresh signal transmissions a dynamic state, and the periodic refresh signal is configured to keep the first communication direction synchronized, and the second direction of communication continues to communicate in an active state.

Asymmetric low-power idle mode can be used for traffic profiling (for example, video surveillance camera links), which can always generate asymmetric transmission scenarios where the network link always transmits data in one direction and the network chain The other direction of the road occasionally transmits a limited amount of data (for example, camera control commands). Since the physical layer device enters the asymmetric low power idle mode without relying on the traffic of the two directions on the network link, the asymmetric low power idle mode adds energy for the network link to generate energy. chance.

On a broad level, the energy-saving control strategy for a specific link in the network determines when to enter the energy-saving state, what energy-saving state to enter (ie, the energy-saving level), how long to maintain the energy-saving state, and energy saving from the previous one. State transitions why energy saving states, etc. In one embodiment, the energy saving control strategy may enable these energy saving decisions based on a combination of settings established by the IT administrator and the link's own traffic characteristics.

Figure 1 illustrates an exemplary link to which an energy saving control strategy can be applied. As shown, the link supports communication between the first link partner 110 and the second link partner 120. In various implementations, link partners 110 and 120 can represent switches, routers, endpoints (eg, servers, clients, VOIP phones, wireless access points, etc.), and the like. As shown, link partner 110 includes physical layer device (PHY) 112, media access control (MAC) 114, and host 116, while link partner 120 includes PHY 122, MAC 124, and host 126.

In general, hosts 116 and 126 may include suitable logic, circuitry, and/or code that may enable the five most recent packets for transmission via a link. The operability and/or functionality of the high functional layer. Since the layers in the OSI model are all served directly by the higher interface layer, the MAC controllers 114 and 124 can provide the necessary services to the hosts 116 and 126, respectively, to ensure that the packets are properly formatted and transmitted to the PHY 112 and 122. The MAC controllers 114 and 124 may include suitable logic, circuitry, and/or code that may enable manipulation of the operability and/or functionality of the data link layer (layer 2). For example, MAC controllers 114 and 124 can be configured to perform an Ethernet protocol, such as an agreement based on the IEEE 802.3 standard. PHYs 112 and 122 may be configured to handle physical layer requirements including, but not limited to, encapsulation, data transfer, and serialization/deserialization (SERDES).

As further shown in FIG. 1, link partners 110 and 120 also include power saving control policy entities 118 and 128, respectively. In general, the energy-saving control policy entities 118 and 128 can be designed to determine when to enter a power-saving state, what energy-saving state to enter (ie, energy-saving level), how long to maintain in the energy-saving state, how to change energy from a previous energy-saving state, etc. .

In general, energy saving control policy entities 118 and 128 may include appropriate logic, circuitry, and/or code that may be enabled to establish and/or implement an energy saving control strategy for a network device. In various embodiments, energy saving control policy entities 118 and 128 may be logical and/or functional blocks, such as may be in more than one layer (including PHY or enhanced PHY, MAC, switches, controllers) Or implemented in other subsystems in the host to enable power saving control in more than one layer.

One feature of the present invention is that an energy efficient Ethernet network, such as the energy efficient Ethernet defined by IEEE 802.3az, can provide significant power savings through the use of an asymmetric low power idle mode. Before describing the details of the asymmetric low power idle mode, a description of the general low power idle mode is first provided.

When the transmitters on both sides of the link enter a silence period when there is no data to transmit, they can enter the general low-power idle mode. In this case, the two transmitters can enter a low power idle mode in which both transmitters are quiet except for short cycle refresh signaling. Figure 2 shows the transfer of refresh signaling at both ends of the link. The use of the low-power idle mode is compared to the transmission of a conventional idle signal when no data is transmitted. It will be appreciated that the transmission of a conventional idle signal will consume exactly as much power as the data transfer.

For link applications such as a one-billion-bit Ethernet network, traffic on either end of the link will require wake-up on both sides of the link. Here, one side of the link will start transmitting data, and the other side of the link will start transmitting idle signals. Figure 3 shows this situation. An example of a traffic profile that can always produce such an asymmetric transfer condition is a video surveillance camera link. Video information is always transmitted in one direction of the video surveillance camera link. In the other direction of the video surveillance camera link, a limited amount of data (eg, camera control commands) is occasionally transmitted. For the latter direction, the idle signal is transmitted for the rest of the time. As this is the case, the presence of data on one end of the link will prevent the other end of the link from entering the low power idle mode.

The inefficiency in this case stems from the inherent two-way protocol where the presence of data on either end of the link will prevent the link from entering a low power idle mode. As shown in Figure 4, data sent from either end of the link does not usually appear at the same time. In other words, there is no correlation between the arrival of data on one end of the link and the arrival of data on the other end of the link. In the current 1000BASE-T specification, for example, an idle signal is transmitted during a period in which data is being transmitted on the other side. In the present invention, it should be appreciated that entering a low power idle mode conditional on the absence of data transfer in both directions of the link limits opportunities for power savings.

In the present invention, it should be appreciated that the asymmetric low power idle mode can create important opportunities for additional power savings. With the asymmetric low-power idle mode, there is traffic in one direction, while the other transmission direction only sends a refresh cycle with quiet periods in between. Figure 5 illustrates the application of an asymmetric low power idle mode to a data pattern similar to that shown in Figure 4. As shown in the figure, energy can be saved even when the flow rate is evenly distributed, because data in one direction may appear in a time interval different from the other direction. Now, these periods in which the idle signal is transmitted in only one direction can be replaced by the refresh signal in the asymmetric low power idle mode. Subsequent to the uninitiated period, replacing the normal idle signal with a periodically transmitted refresh signal represents additional power savings on the network link.

As will be appreciated, the additional benefit provided by the asymmetric low power idle mode is more pronounced where these flows occur primarily in one direction. Figure 6 illustrates this flow representation of an asymmetric traffic pattern, such as a video surveillance link. As shown in the figure, while the continuous data flow is being transmitted in the other direction, the low power idle mode can be entered in only one direction. This is in turn different from the transmission of the normal idle signal in the example shown in FIG.

In the asymmetric low power idle mode, two additional low power idle states can be defined. In the first low power idle state, the transmitter is enabled and the receiver is in a low power idle mode, while in the second low power idle state, the receiver is enabled and the transmitter is in a low power idle mode. This is in contrast to the traditional low-power idle mode where both the transmitter and receiver are active or both the transmitter and receiver are in a low power idle mode.

In the present invention, it will be appreciated that for digital signal processing (DSP) blocks, there are new challenges to enabling these two new asymmetric low power idle mode states. For example, consider that the receiver is started and the transmitter is in low power idle mode. Asymmetric low power idle state. Here, when the data is received, turning off the sender will affect the echo/next response. Thus, in one embodiment, the transmitter can be configured to transmit zero until the echo/canceler buffer is filled with all zeros before being turned off. Sending zeros during wake-up of the transmitter can also be used to avoid the initial unstable transition state of the transmitter.

In another example, consider an asymmetric low power idle state where the transmitter is enabled and the receiver is in a low power idle mode. Here, when the transmitter is activated, the received echo may be large enough to trigger the analog signal detector. Thus, in one embodiment, the signal level can be checked after the echo/next cancellation for receiving signal detection. Therefore, it may be necessary to manage DSP self-tuning based on responses in the startup and stabilization cycles.

Although not shown in the figure, there is a sleep cycle when the PHY transitions to the low power idle mode. This sleep cycle may be approximately 200 μs. For Gigabit Ethernet, even if a very short packet needs to be sent, the boot cycle exceeds 216 μs (ie, sleep time (Ts) plus wake-up time (Tw). To this end, wake up one direction for easy transfer of small amounts of data. This means significant inefficiency. For traffic profiles that include a limited amount of data being transmitted in infrequent intervals, this possible factor can significantly affect the level of energy savings available to the network link.

In one embodiment, the refresh period in the asymmetric low power idle mode can be used for the transmission of a limited amount of data. In one embodiment, the PHY may send a signal to the MAC (eg, using a combination of RX_ER, RX_DV, and RXD signals defined by the MAC interface) to inform the PHY that it is in a refresh cycle and is ready to transmit data. The MAC can then choose to use this refresh cycle for data types that cannot be delayed in sensitivity. In various examples, the data types that may utilize the refresh cycle may be higher layer network management information packets, control commands (eg, video camera commands), uplink data during internet browsing, Wait. As will be appreciated, the specific mechanism by which the refresh period can transfer a limited amount of data can be implemented by attachment. For example, the refresh cycle can convey a finite amount of data that can be identified with a single refresh cycle sequence or more than two refresh cycle sequences.

In general, the refresh cycle is capable of establishing a fixed lower data rate logical channel at the MAC using the PHY refresh cycle. In the case of a Gigabit Ethernet network, the refresh cycle can provide a slower logical link of up to 10 Mbps with a refresh cycle. Significantly, building a slower data link eliminates the need to wake up the PHY. This facilitates higher power savings by enabling transmission in one direction to remain in a low power idle mode.

In one embodiment, the PHY may include a buffer that enables continuous slower links at the MAC while utilizing a refresh cycle for transmitting data. When in this mode, the clock between the PHY and the MAC may be slower.

In general, asymmetric low power idle can reduce power consumption and emissions when applied to asymmetric or other uncorrelated data flow modes. The advantage of asymmetric low-power idle is to increase the chance of reducing power consumption in the transmitter or receiver.

The asymmetric low power idle mode has been described above, and now with reference to the flow chart of Figure 7, it shows a process flow diagram using an asymmetric low power idle mode. As shown, the process begins in step 702 where the network device begins to operate in an active mode in which transmissions in both directions on the network link operate at a defined data transmission rate. For example, the 1000BASE-T PHY operates at a data transfer rate of 1 Gb/s in both directions.

In step 704, while the network device continues to operate in the startup mode, the network device will then monitor the link utilization level of the first communication direction. In one embodiment, the energy saving control policy implemented in the network device Perform monitoring a little. As can be appreciated, the energy saving control strategy can be implemented in more than one layer of the network device. Here, it should be noted that the monitoring of the first communication direction shown in FIG. 7 does not mean that the monitoring of the second communication direction is excluded. Instead, the example of Figure 7 is provided to illustrate the asymmetric process.

In step 706, it is then determined if the link utilization level falls below a threshold. As will be appreciated, the particular type of threshold will depend on the type of indicator used to determine the link utilization level. In one example, the link utilization level can be determined based on a traffic sequence or buffer level, more than one device or subsystem state, an application activity, and the like. Regardless of the type of indicator used, monitoring of the link utilization associated with the threshold level can be used to determine if the network device can enter the asymmetric low power idle mode.

Specifically, if it is determined in step 706 that the monitored link utilization level in the first communication direction has not fallen below the threshold, the process will continue to operate in the startup state and monitoring will continue in step 704. On the other hand, if it is determined in step 706 that the monitored link utilization bit in the first communication direction has been accurately below the threshold level, the process proceeds to step 708, in which the first communication direction can be converted to low power consumption. Idle mode.

In this process, the energy saving control strategy can generate a control signal that will indicate that the component in one direction of transmission enters a low power idle mode. For example, the power saving control strategy may generate a control signal indicating that the transmitting subsystem enters a low power idle mode, or may generate a control signal indicating that the receiving subsystem enters a low power idle mode. In one embodiment, the control signal for transitioning to an asymmetric low power idle mode may be generated by the MAC. In another embodiment, a control signal for transitioning to an asymmetric low power idle mode can be generated within the PHY that analyzes the traffic received from the MAC.

As described above, the asymmetric low power idle mode can enhance power savings by eliminating the barrier to entry into the low power idle mode due to un-started requirements in both communication directions.

Another embodiment of the present invention may provide machine and/or computer readable memory and/or media having stored thereon machine code and/or computer having at least one portion of code executable by a machine and/or computer. Program to boot the machine and/or computer to perform the steps described herein.

These and other aspects of the invention will be apparent to those skilled in the < Although a few features of the invention have been described above, the invention may be embodied and carried out in various ways that are apparent to those of ordinary skill in the art after reading the invention. Therefore, the above description should not be considered as excluding those other embodiments. Further, it is to be understood that the phraseology and terminology used herein are for the purpose of description

110‧‧‧First Link Partner

112‧‧‧PHY

114‧‧‧MAC

116‧‧‧Host

118‧‧‧Energy Conservation Control Strategy Entity

120‧‧‧second link partner

122‧‧‧PHY

124‧‧‧MAC

126‧‧‧Host

128‧‧‧Energy Conservation Control Strategy Entity

702~708‧‧‧Steps

Figure 1 shows an Ethernet link between link partners.

Figure 2 shows the refresh signaling between link partners in a low power idle mode.

Figure 3 shows signaling with one-way data between link partners.

Figure 4 illustrates the use of asymmetric low power idle for asymmetric data traffic.

Figures 5 and 6 illustrate the use of asymmetric low power idle for asymmetric data traffic.

Figure 7 shows a process flow diagram of the present invention.

110‧‧‧First Link Partner

112‧‧‧PHY

114‧‧‧MAC

116‧‧‧Host

118‧‧‧Energy Conservation Control Strategy Entity

120‧‧‧second link partner

122‧‧‧PHY

124‧‧‧MAC

126‧‧‧Host

128‧‧‧Energy Conservation Control Strategy Entity

Claims (15)

An energy saving method comprising: operating a network device in a startup mode, in which the physical layer device and the link partner in the network device are in a communication direction of 1 Gb/s via a network link Transmission rate communication; monitoring, by the network device, a link utilization level with respect to the first communication direction via the network link; and in response to the monitoring, changing the physical layer device from the startup mode to the non-transmission mode a symmetric low power idle mode, wherein the asymmetric low power idle mode supports configuration of the physical layer device, wherein the first communication direction is configured to maintain a synchronization period of the first communication direction The transmission of the sexual refresh signal is in an unactivated state, and the second communication direction opposite to the first communication direction continues to communicate at the transmission rate of 1 Gb/s; wherein the first communication direction is used The periodic refresh signal initiates data communication in the asymmetric low power idle mode to establish a fixed data rate logical channel, the fixed data rate logical channel having data below 1 Gb/s Transmission rate. The energy saving method of claim 1, wherein the first transmission direction is transmitted from the network device to the link partner. The energy saving method of claim 1, wherein the first transmission direction is transmitted from the link partner to the network device. The energy saving method according to claim 1, wherein the physical layer device is a 1000BASE-T physical layer device. The energy saving method of claim 1, wherein the network link is a video surveillance camera link. The energy saving method according to claim 1, wherein the fixing The data rate logic channel has a data rate of 10 Mb/s. An energy saving method comprising: operating a network device in an active mode, in which the physical layer device in the network device supports and the link partner is bidirectional at a defined data transmission rate via a network link Communicating; and responding to the link utilizing level analysis to transition the physical layer device from the boot mode to an asymmetric low power idle mode, wherein the asymmetric low power idle mode supports the physical layer device Configuration, wherein the first communication direction operates at the defined data transmission rate, and the second communication direction is not activated between transmissions of the periodic refresh signal configured to maintain synchronization of the first communication direction, The first communication direction uses the periodic refresh signal to initiate data communication in the asymmetric low power idle mode to establish a fixed data rate logical channel to establish a maximum data having a lower than the defined data transmission rate. Rate of fixed data rate logic channels. The energy saving method of claim 7, wherein the first transmission direction is transmitted from the network device to the link partner. The energy saving method of claim 7, wherein the first transmission direction is transmitted from the link partner to the network device. The energy saving method according to claim 7, wherein the physical layer device is a 1000BASE-T physical layer device. The energy saving method of claim 7, wherein the network link is a video surveillance camera link. The energy saving method according to claim 7, wherein the logical channel is a fixed data rate logical channel having a data rate of 10 Mb/s. A network device comprising: a transmitter configured to transmit to the link partner at a transmission rate of 1 Gb/s via a network link; a receiver configured to transmit from the chain at the transmission rate of the 1 Gb/s via the network link And the energy saving control strategy module is configured to analyze a link utilization level of the first communication direction via a network link connecting the network device and the link partner device, the energy saving control strategy The module is further configured to, in response to the analyzing, transition the network device to an asymmetric low power idle mode, wherein the asymmetric low power idle mode supports the transmitter to periodically refresh the signal Configuring between the transmissions as an inactive state and configuring the receiver to receive data at the 1 Gb/s transmission rate, the periodic refresh signal being configured to remain synchronized with the link partner device; The transmitter initiates data communication in the asymmetric low power idle mode by using the periodic refresh signal to establish a fixed data rate logical channel, and the fixed data rate logical channel has a resource lower than 1 Gb/s. Transmission rate. The network device of claim 13, wherein the transmitter and the receiver are part of a 1000BASE-T physical layer device. The network device of claim 13, wherein the fixed data rate logical channel has a data rate of 10 Mb/s.